The validity of attribute-importance measurement: A review

A critical review of the literature demonstrates a lack of validity among the ten most common methods for measuring the importance of attributes in behavioral sciences. The authors argue that one of the key determinants of this lack of validity is the multi-dimensionality of attribute importance. Building on the notable work of Myers and Alpert (1968) [Myers JH, Alpert MI. Determinant buying attitudes: Meaning and measurement. J Mark 1968;32(July):13–20], they propose that different methods measure different dimensions of attribute importance and, more specifically, what methods measure which specific dimensions. A re-examination of existing research reveals convergent and nomological validity among methods that are proposed to measure the same dimensions of attribute importance and discriminant validity between methods that are proposed to measure different dimensions of attribute importance. Acknowledging the multi-dimensionality of attribute importance substantially reduces the apparent lack of validity reported in the literature and forms an important first step enabling practitioners and scholars to improve the validity of attribute-importance measurement.